Re: [RFC] Scalable statistics counters using kmalloc_percpu

[Andrew Morton <akpm@zip.com.au>]

Ravikiran G Thirumalai wrote:
> 
> Here is a Scalable statistics counter implementation which works on top
> of the kmalloc_percpu dynamic allocator published by Dipankar.
> This patch is against 2.5.27.
> 
> ...
> +static inline int __statctr_init(statctr_t *stctr)
> +{
> +       stctr->ctr = kmalloc_percpu(sizeof(*(stctr->ctr)), GFP_ATOMIC);
> +       if(!stctr->ctr)
> +               return -1;
> +       return 0;
> +}

Minor nit: please force the caller to pass in the gfp_flags when
designing an API like this.  The fact that you were forced to use
GFP_ATOMIC here shows why...

> +       for( i=0; i < NR_CPUS; i++ )
> +               res += *per_cpu_ptr(stctr->ctr, i);
> +       return res;
> +}

Oh dear.  Most people only have two CPUs.

Rusty, can we *please* fix this?  Really soon?


General comment:  we need to clean up the kernel_stat stuff.  We
cannot just make it per-cpu because it is 32k in size already.  I
would suggest that we should break out the disk accounting and
make the rest of kernel_stat per CPU.

That would be a great application of your interface, and a good
way to get your interface merged ;)  Is that something which you
have time to do?

Thanks.